What is tokenlean CLAUDE.md?

Project instructions for tokenlean, a command-line toolkit that helps AI agents inspect projects while using fewer tokens. They describe its folders, shared modules, conventions, and process for adding tools.

Use them when adding a tokenlean command, updating its Node.js package configuration, or following its shell, output, and project-analysis conventions.

They give contributors a consistent way to extend the command-line tools without duplicating formatting, traversal, caching, or safety logic.

tokenlean CLAUDE.md — Instructions file by edimuj

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods

npx agentmods add instructions/edimuj/tokenlean/claude-md

Clone the repo

git clone --depth 1 https://github.com/edimuj/tokenlean

Per session 583 This file is loaded in full into every session.

When invoked 583 The same file — it is already loaded in full.

Security scan A 1 finding. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5	$0.00583	$0.00583
Opus 5	$0.00292	$0.00292
Sonnet 5	$0.00117	$0.00117
Haiku 4.5	$0.00058	$0.00058

Measured 2d ago against content hash 138012bb263c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

tokenlean CLAUDE.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Runs shell commandslowCapability

Expected in a hook, worth knowing in a rule or an instructions file.

shell.mjs # Safe git/rg wrappers (spawnSync, no shell injection)

CLAUDE.md · 60 lines

How it starts

The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Tokenlean

CLI tools for AI agents - maximum insight with minimum tokens.

Project Structure

bin/               # CLI entry points (tl-*.mjs) - thin wrappers
src/               # Shared modules
  output.mjs       # Common flags (-l, -t, -j, -q, -h), formatting, token estimation
  project.mjs      # File categorization, language detection, skip/important logic
  cache.mjs        # LRU disk cache for expensive operations (ripgrep results)
  traverse.mjs     # Fast file traversal with filtering
  config.mjs       # Config file loading (.tokenleanrc.json)
  shell.mjs        # Safe git/rg wrappers (spawnSync, no shell injection)
  generic-lang.mjs # Regex-based symbol/import extraction for non-JS/TS languages
skills/            # Agent skills
  claude/          # Claude Code skill variants
  codex/           # Codex skill variants

Adding a New Tool

Create bin/tl-<name>.mjs with shebang #!/usr/bin/env node
Import shared modules from ../src/
Use output.mjs for consistent flags and formatting
Add entry to package.json bin section
Run npm link to test locally

Conventions

Minimal dependencies - node-html-markdown for tl-browse, otherwise Node.js built-ins only
ES modules - use .mjs extension, import/export
All tools support common flags: -l (limit), -t (tokens), -j (JSON), -q (quiet), -h (help)
Users invoke tools via tl <command> (e.g. tl symbols, tl run), which dispatches to bin/tl-<command>.mjs
For noisy logs, prefer tl tail <file> (or piped stdin) to collapse repeats and surface error/warn lines
Exit codes: 0 success, 1 error
Output should be token-efficient (no verbose explanations)

Shell Completions

tl completions bash|zsh outputs a completion script. Dynamic — new tools are picked up automatically.

Backend: tl --list-commands [--with-desc] and tl --list-flags <command> (extracts flags from tool source via regex).

Testing

npm link              # Install locally
tl <tool> --help      # Verify help works
tl <tool> [args]      # Test functionality

Read the full file on GitHub · 60 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

2d ago First seen · 60 lines · 583 tokens per session scan A 138012bb263c

Subscribe to this mod's changes

tokenlean CLAUDE.md is an instructions file published in the GitHub repository edimuj/tokenlean (11 stars, last pushed 1mo ago), licensed MIT. It adds 583 tokens to every session, about $0.0029 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

agent-tools ai-agent-tools cli-tools context7

h5i-ctx2img AGENTS.md

Instructions for Koukyosyumei/h5i-ctx2img, covering persona, h5i integration, workflow, code commits and capturing large command output (token reduction).

Koukyosyumei/h5i-ctx2img · 1,313 tokens

h5i-ctx2img CLAUDE.md

Instructions for Koukyosyumei/h5i-ctx2img, a project described as: Render agent context as images. Same text, 60–75% fewer tokens.

Koukyosyumei/h5i-ctx2img · 15 tokens

stealth-browser-mcp AGENTS.md

Instructions for vibheksoni/stealth-browser-mcp, covering agent instructions, start here, recommended proxy services, environment and validation.

vibheksoni/stealth-browser-mcp · 712 tokens

mcp-steroid CLAUDE.md

Instructions for jonnyzzz/mcp-steroid, covering claude.md, agents.md, design philosophy, recursive context lookup (do this before sub-folder work), sub-folder guides and must do.

jonnyzzz/mcp-steroid · 6,916 tokens

untether GEMINI.md

Instructions for littlebearapps/untether, covering untether — gemini instructions, stack & conventions, architecture, key rules and testing.

littlebearapps/untether · 500 tokens

kinocut copilot-instructions.md

Copilot instructions for KyaniteLabs/kinocut: Read the nearest AGENTS.md for repository rules. For Kinocut workflows, load skills/kinocut/SKILL.md; it is the canonical public skill. Keep this adapter limited to host-specific discovery and do not duplicate policy here.

KyaniteLabs/kinocut · 60 tokens

tokenlean CLAUDE.md